news and views
Clearing a path for nerve growth
Lars Olson
After spinal-cord injury, severed nerve fibres face a thicket of obstacles as
they try to regenerate. Researchers have used a bacterial enzyme to help
prune these obstacles in rats.
Cell
b
Tip of nerve fibre with growth
cone and exploring filopodia
Space
between
cells
Chondroitinase
ABC prunes CSPG
P
STO
NATURE | VOL 416 | 11 APRIL 2002 | www.nature.com
a
STOP
S
STOP
pinal-cord damage is tragically common, affecting three to five people out
of every 100,000 in the United States
and similar numbers in other countries.
The consequences for both the injured
person and their families and friends can
be devastating. Injuries typically occur in
young adults, and of the 11,000 or so new
cases each year in the United States, 80% are
men who were involved in driving accidents,
violence or falls. Most victims survive, but
are completely or partially paralysed for
the rest of their lives.
It has long been thought that such traumatic injuries are incurable, but intense
experimental work is beginning to suggest
that they might one day be treatable at least.
In the latest advance, Bradbury and colleagues1 provide evidence (page 636 of this
issue) that infusion of chondroitinase ABC
— a bacterial enzyme that trims the carbohydrate side chains off large extracellular
proteins — enables severed nerve fibres to
regenerate after the spinal cord has been
crushed in rats. The authors used forceps to
squeeze the nerve fibres in the so-called dorsal columns of the spinal cord until nerve
fibres carrying sensory information from the
limbs to the central nervous system, and
fibres conveying motor (movement) information to the limbs, had been cut. This
severely impairs limb function. In animals
that were then treated with chondroitinase
ABC, there was convincing evidence of
nerve-fibre regeneration at the cut nerve
stumps. Although the growth of sensory
nerve fibres was not enough to lead to a clear
recovery of sensation, the authors have electrophysiological and behavioural evidence
of remarkable motor recovery.
Imagine that you are as small as the tip
of a nerve fibre, a few thousandths of a millimetre in diameter, and have to advance in the
narrow regions between cells — the extracellular space — in a recently damaged tissue. You would find that these spaces are by
no means open territory, but rather like an
endless overgrown shrubbery, consisting
of many different long, and sometimes
branched, molecules that form an intertwined network, the extracellular matrix
(Fig. 1). Some of the molecules are anchored
to cells, some hold adjacent cells together,
and many are simply deposited outside the
cells. The terrain becomes particularly hostile after injury because new classes of cells
STOP
STOP
STOP
Cell
STOP
Cell-membrane-anchored
molecules that inhibit nerve-fibre
growth (best known: Nogo)
Chondroitinase ABC
Antibodies
CSPG molecules
Figure 1 Obstacles to nerve regeneration. a, A severed nerve is shown on the left, in the extracellular
space between cells in a wounded spinal cord. The nerve end has just formed a ‘growth cone’, which
is exploring the territory by constantly forming and retracting minute processes (filopodia). The
extracellular space contains two obstacles to nerve growth: chondroitin sulphate proteoglycan
(CSPG) molecules, which have a backbone and many side branches that block the way; and special
cell-membrane-anchored proteins that actively stop growing nerve fibres. The best known of these
is Nogo. b, Bradbury et al.1 show that the bacterial enzyme chondroitinase ABC can prune the side
chains of CSPGs, clearing the way for growing nerve fibres. It has previously been shown that Nogo
can be blocked by specific antibodies. The antibodies and enzyme could potentially be used in
combination, along with proteins that stimulate nerve growth.
arrive, and a scar that is burdened by increasing deposition of matrix molecules builds
up2,3. To advance, you would need a machete.
The chemistry of the extracellular matrix
is complex, but one key class of molecules are
the chondroitin sulphate proteoglycans
(CSPGs). These are made up of a protein
core — the chondroitin sulphate — that is
equipped with many side chains consisting of carbohydrate building-blocks such
as glycosaminoglycans. Bacteria such as
Proteus vulgaris have evolved enzymes,
including chondroitinase ABC, that can
prune CSPGs by removing glycosaminoglycans. Perhaps this helps the bacteria to
invade their animal hosts.
Knowing all this, Bradbury et al.1 wondered whether chrondroitinase ABC could be
used as a molecular machete to clear the way
for nerve regeneration in injured rat spinal
cords. After crushing the dorsal columns of
rats at the level of the fourth cervical vertebra,
the authors repeatedly infused either active
chondroitinase ABC or a control protein,
delivered through a thin tube, to the site of
injury. Then, using an antibody that binds to
the CSPG core but not the intact molecule,
they showed that chondroitinase ABC effec© 2002 Macmillan Magazines Ltd
tively prunes CSPGs in the spinal cord
(Fig. 1). CSPGs remained intact in controls.
In the treated rats, sensory neurons in the
dorsal root ganglia corresponding to the
fourth cervical vertebra upregulated a nervegrowth marker protein, GAP-43, and the cut
extensions (axons) of these neurons in the
spinal cord regenerated by up to 4 millimetres towards the brain1. Similarly, the important dorsal corticospinal tract, completely
severed by the crush, regenerated downwards from the site of injury in the spinal
cord. This was parallelled by a return of electrical signalling between the brain’s cerebral
cortex and the spinal cord, albeit at a reduced
strength and somewhat slower speed. Bradbury et al. also showed that if they cut the
spinal cord again, stimulation of the cerebral
cortex no longer elicited signals in the cord.
In attempting to repair experimentally
damaged spinal cords, it is crucial to see
how well the animals recover. Bradbury
et al. went to great lengths to test this.
They show that rats that had been treated
with chondroitinase ABC recovered normal or near-normal walking behaviour in
two tests — beam and grid walking. But
sensorimotor functions — becoming aware
589
news and views
of and removing a piece of adhesive tape —
were not significantly recovered.
The results1 are promising nonetheless,
and it might prove possible to integrate the
use of chondroitinase ABC with other strategies (compare with ref. 4). Let’s consider the
worst-case scenario, in which the spinal cord
has become completely divided, leaving a
gap between two stumps. While scar tissue
invades the gap, initial oedemas (swellings)
in the stumps themselves may transform
into growing, fluid-filled cysts5. Yet both
stumps contain severed axons with an intrinsic ability to regenerate — if the conditions
are right. The problem can be likened to
opening up traffic on a blocked-off, overgrown and partly demolished road. We may
need to remove scar tissue, empty cysts, build
physical bridges such as nerve grafts across a
void6, and perhaps provide cellular surfaces
that prompt nerves to grow7–9.
To clear the road, chondroitinase ABC
could be delivered to prune back the extracellular-matrix shrubbery. Chemical stop
signs — such as those composed of the
nerve-growth-inhibitory protein Nogo,
expressed by supportive glial cells — could
be covered over by antibodies and thus neutralized10. Chemical gradients of nervegrowth-activating molecules may need to be
established to stimulate and to direct growing nerve fibres. Finally, to protect newly
regenerated nerve fibres and increase signalling speed, cells that provide fibres with a
new insulating sheath (composed of a lipid,
myelin) might be needed. It may prove easier
to induce long axon growth in white matter
than in grey matter. Grey matter is made up
mainly of nerve-cell bodies, with some other
cells, and the complex network of contacts
between nerve-fibre branches. White matter
consists largely of myelin-coated nerve
fibres. Any fibres in grey matter tend to
branch greatly, so the neuron’s need for
growth factors from underlying cells may be
satisfied long before it reaches its target area.
But fibres in white matter branch less, so
can perhaps grow further before they are
adequately supported.
What could go wrong by using chondroitinase ABC? Not much, it seems. Highly
purified preparations have no proteindegrading activity and so would not destroy
crucial proteins in the body. However,
CSPGs come in many forms, and some
might have specific roles in the central nervous system, for example to guide growing
axons. In grey matter, many nerve-cell
bodies are surrounded by nets of particular
CSPGs, which might be important in
neuronal contacts. These nets would
degrade upon enzyme treatment, and take a
long time to rebuild.
That said, treatment with chondroitinase
ABC has also proved effective in treating
brain injuries11, and constitutes one hopeful
avenue to promoting nerve regeneration.
590
Several other treatment strategies (see, for
example, ref. 12) have also shown promise
in different experimental models of spinalcord injury. The bad news is that none of
these techniques has led to complete recovery. Fortunately, however, many of them can
be combined, perhaps in ways discussed
above. In the short term, the prognosis for
people with complete spinal-cord injury
remains grim. Yet, looking further into the
future, we can perhaps allow ourselves to be a
bit more optimistic.
■
Lars Olson is in the Department of Neuroscience,
Karolinska Institute, 171 77 Stockholm, Sweden.
e-mail: [email protected]
1. Bradbury, E. J. et al. Nature 416, 636–640 (2002).
2. Lemons, M. L., Howland, D. R. & Anderson, D. K. Exp. Neurol.
160, 51–65 (1999).
3. Davies, S. J. et al. Nature 390, 680–683 (1997).
4. Olson, L. Nature Med. 3, 1329–1335 (1997).
5. Josephson, A. et al. Neurosurgery 48, 636–645 (2001).
6. Cheng, H., Cao, Y. & Olson, L. Science 273, 510–513 (1996).
7. Li, Y., Field, P. M. & Raisman, G. Science 277, 2000–2002
(1997).
8. Ramon-Cueto, A. et al. Neuron 25, 425–435 (2000).
9. Hofstetter, C. P. et al. Proc. Natl Acad. Sci. USA 99, 2199–2204
(2002).
10. Fouad, K., Dietz, V. & Schwab, M. E. Brain Res. Brain Res. Rev.
36, 204–212 (2001).
11. Moon, L. D. F. et al. Nature Neurosci. 4, 465–466 (2001).
12. Coumans, J. V. et al. J. Neurosci. 21, 9334–9344 (2001).
Plant biology
The first harvest of crop genes
Michael Bevan
Draft sequences of the rice genome have been produced by two groups.
The drafts will be an invaluable resource for research on the genomes of
other plants, the cereals in particular.
ast week’s publication in Science1,2 of two
draft sequences of the rice genome is
another milestone in the inexorable
progress of genome sequencing. The work is
significant on three counts. Rice (Oryza sativa) is the staple in the diet of nearly a third
of the world’s population; the sequences
provide a reference for all the other cereals,
which collectively account for an overwhelming proportion of human and domesticated
animal feed; and this is the first representative
of the monocotyledonous plant lineage to be
sequenced. Together with the dicotyledonous
plants, the monocots (grasses, palms, rushes,
and so on) constitute the flowering plants, the
most diverse and numerous of all plants.
One of the draft sequences has been produced by Syngenta (a Swiss agricultural
biotech company)1; the other comes from a
collaboration between the Beijing Genomics
Institute and other Chinese academies2. Both
reports are initial analyses of assembled
whole-genome shotgun sequences, involving
random sequencing of the entire genome and
assembly of the sequence into contiguous
regions, or contigs. The Syngenta group
worked with the subspecies japonica Nipponbare (a popular rice cultivar in Japan),
whereas the Chinese team sequenced a
hybrid indica variety, the major subspecies
grown in China and most other regions.
Draft DNA sequence is generally accurate
enough to identify genes and to estimate
genome size. In this case, the sequences reveal
a genome of 362–389 million base pairs
(Mbp), and a conservative number of
between 33,000 and 44,000 genes. Two other
important studies of the rice genome provide
independent evidence, based on the construction of a near-complete physical map3
and a map of expressed genes4, that the gene-
L
© 2002 Macmillan Magazines Ltd
Monocots
Species
Rice
Maize
Barley
Wheat
EST Genomic
Tomato
Potato
Dicots
Medicago
Lotus
Soybean
Arabidopsis
Brassica
Poplar
Conifers
Mosses
Pine
Spruce
Physcomitrella
Figure 1 Progress in plant-genome sequencing.
Large-scale sequencing of representatives of
several of the major groups of plants is
underway (red bars). Mostly this is through
analysis of expressed sequence tags (ESTs),
as with pine and moss (Physcomitrella). But
genome sequencing of Brassica species and
Lotus (a model legume) is ongoing, with various
approaches being used, and work on maize,
Medicago (alfalfa) and poplar will start
imminently (pale red bars). The evolutionary
relationships shown here are figurative only.
rich regions of the rice genome total about
400 Mbp and contain about 44,700 genes.
The rice genome is much smaller than
that of maize (which, at 2,500 Mbp, is about
the same size as the human genome), and is
tiny compared to the monster barley (twice
NATURE | VOL 416 | 11 APRIL 2002 | www.nature.com